Lasting Partnerships that move new and innovative ideas into the market

123
123

Lasting Partnerships that move new and innovative ideas into the market

Lanza techVentures invests in broad, forward-looking areas such as health and healthcare, pharma, computers, community-based, open-source silicon platforms, design tools and the intersection of AI and medical

For your consideration…

Looking back, the impact of the Internet on society is similar to the agricultural society moving into the industrial age. When you consider the impact of 300-milion individuals connected in the U.S. to two-billion others around the world, you see its strong impact on society.

The traditional chip design process won’t work moving forward. Instead, it will be based on silicon IP that can be used with minimal cost because a chip today should be able to be designed quickly, cheaply and boldly by anyone with a natural ability.

Open source for a creative environment and strategic advantage is the way to go because it offers the largest set of low-cost implementation tools. New communities of designers, in fact, are working on new generations of design using this approach.

Automated design learning with a growing brain and memory is improving constantly and is the next step of the industrial revolution. Automated AI is five to 20 years away and will impact society with infinite compute power. Medical information, for example, will be reevaluated and the next chip to design itself will do so automatically. It will know which processes are available and be able to deliver a prototype as it keeps learning.

In the past, semiconductors were designed all the way through manufacturing. It was heavy-duty deep silicon. Next came semi-custom design and then silicon IP. What’s next? What will it take to get an AI chip into full production? On the other side of the equalization, how will verification experts respond to the explosion of silicon-on-package designs that may be low end today but could develop into broadly used deep-silicon designs?

IoT and chiplets are the next semi-custom design. Libraries of IP are the best way to create chiplets for lower cost with more designs in less time.

What other people see as a mishmash, I see as the center of the bull’s eye. A chip today should be able to be designed by anybody quickly, cheaply and boldly.